The University of Chicago Magazine June 1995
The facts at hand: Kris Hammond's new program puts Internet wisdom within reach.
IF YOU WANT ANSWERS in a hurry, the Internet's broad resources are often no better than an electronic Tower of Babel. Tools like Gopher servers, the World Wide Web, and the public bulletin boards called newsgroups organize information by totally different schemes, and there's no comprehensive index. Veterans users search the "Net" by browsing, but, says Chicago computer scientist Kris Hammond, that's of little help to most people.
Still, small pockets of cyberspace do have a predictable structure, allowing quick access to helpful data. For Hammond, associate professor and director of the Artificial Intelligence Laboratory, these regions are allowing him and his colleagues to place a part of the Internet's resources within reach of a much wider audience. Their project is a program called FAQ Finder, named for the files of "frequently asked questions" compiled by the thousands of Internet newsgroups -- electronic forums where people with a common interest post ideas and ask questions.
A newsgroup's FAQ file, explains Hammond, orients new subscribers who, like visitors to any information center, often ask the same questions time after time. Someone reading a newsgroup about the movies "will immediately ask, `What does "Rosebud" mean in Citizen Kane?' `What does a gaffer do?' Veterans don't want to answer every time."
Hammond -- with visiting assistant professor Robin Burke, Argonne researcher Terry Gaasterland, and DePaul University colleague Steve Lytinen -- has turned the FAQ files of far-flung newsgroups into a single, practical database containing devotees' wisdom on anything from movies to mathematics, from the Middle East to martial arts.
Recycling that knowledge exemplifies a strategy that Hammond pioneered in the field of artificial intelligence -- case-based reasoning. It's a tactic, he says, that humans use all the time: "When you look at a new problem, you try to find an existing problem in memory that is of the same kind."
But FAQ Finder is more than an index of old ideas. It seeks out appropriate answers starting with a user's question -- typed in plain English. If any "frequently asked question" matches the user's original query, it displays the question -- and the accompanying answer. The use of a sophisticated semantic analysis and an electronic dictionary and thesaurus allow for a seemingly human response: A query like "How do I separate my husband's credit rating from mine?" is matched in under a minute to the FAQ, "How do I get my spouse off my credit rating?" Most "intelligent" programs that search the Internet, says Hammond, look for text without analyzing its meaning. "What happens is, if you're interested in snakes and say, `Give me files with Python in it,' " you'll also get unwanted files on the comedy troupe Monty Python.
Hammond's group plans to release FAQ Finder as a free, public resource, and reference librarians and Internet users will no doubt find it a boon. (The program will be available on the World Wide Web by late summer at http://cs-www.uchicago.edu/ ~burke/faqfinder.html.) The Office of Naval Research is funding a classified application of the program, and Hammond is optimistic about finding commercial licenses -- adapting the program, perhaps, as a search tool for an electronic publication, or for use by product- or service-help desks.
Given the Internet's crazy quilt of organizational schemes, though, he's less sanguine about expanding FAQ Finder to comb more of cyberspace. Other programs' simple word searches are one thing, but, he says, "right now, it's not feasible to build anything even slightly intelligent that would cruise the Net."
Location, location: Towns like Morris, shown in 1881, thrived thanks to the Illinois & Michigan Canal.
EXPLORER LOUIS JOLLIET POINTED it out in 1673: If not for a single, 100-mile stretch of land, people and goods could travel easily by boat from the Great Lakes to the Gulf of Mexico. In 1848 his vision was realized when the Illinois & Michigan Canal opened, joining Lake Michigan to the Illinois River and, through it, to the Mississippi and the Gulf. The I & M Canal, says historical geographer Michael Conzen, was America's "last big canal that made a difference. The difference it made was Chicago."
Today, the region along the former canal is an outdoor laboratory for Conzen's College and graduate students. A professor and chair of the committee on geographical studies, he uses the landscape and towns to illustrate the role geography has played in American history.
In the case of the I & M Canal, which stretched west from Chicago to LaSalle-Peru, that role was enormous. "Most people think the railroads catapulted Chicago," says Conzen, "but they came later." In the canal's first six years, an industrial corridor sprang up along the waterway, which "articulated the whole upper Midwest as a region." Trade that had flowed south to New Orleans now went east to Chicago -- a trend only reinforced, he explains, when the railroads arrived.
Dotted with small towns every few miles, the canal region is rich in history. "Thoreau, Mark Twain -- all these luminaries came through the I & M Canal corridor on speaking tours," notes Conzen. "Ottawa is the site of the first Lincoln-Douglas debate." His students explore the area through research projects chosen from a list that Conzen prepares, each year focusing on different towns -- Joliet or Lockport, for example -- or issues, such as settlement patterns or the industrial revolution. The corridor's proximity to Chicago, he says, creates an "extraordinary" teaching resource. "We can easily drive there to explore archives, to meet people, to collect information, to look at the landscape."
The benefits flow two ways. For the past eight years, Conzen has edited and published the best of each class's research papers in a series, Studies on the Illinois & Michigan Canal Corridor. What is usually the students' first scholarly publication is often the first geographical survey of a particular town or regional historical development -- an information gap that prompted the series after local librarians "begged" for copies of the students' final papers.
In the years since the I & M Canal closed in 1933 -- it was largely outmoded when the much wider Chicago Sanitary & Ship Canal opened in 1900, and some sections are today completely dry -- many nearby factory jobs disappeared. But, in 1984, the region received a boost when it was designated the country's first National Heritage Corridor, arousing interest in preserving its history and in how, through tourism, that history could help local economies.
Conzen, who assisted the groups that lobbied for National Heritage Corridor status, says that these trends have made his students' research more and more valuable. Already the papers have been cited in both scholarly footnotes and efforts to place sites on the National Register of Historic Places. Before Conzen began the series, he says, "The students wrote papers and they ended up in my file cabinet." Now, the work has found its own sense of place.
Mystery killer: Why did the event that wiped out Albertosaurus and other dinosaurs spare snails like Cerithiella?
Pity the poor dinosaurs. Falling asteroids, lava-spewing volcanoes, rising sea levels, and global cooling: In the collective notebooks of scientists, T. rex and company have suffered their share of calamities. What force really killed off the dinosaurs 65 million years ago? The popular whodunit is part of the larger mystery of mass extinctions, the cataclysmic events that radically changed the course of evolution at least five times. Their case is far from solved, but, in a trio of recent studies, Chicago researchers have turned up several new leads, possibly ruling out some suspects in the mass extinction that wiped out dinosaurs and countless other species.
What David Jablonski calls a "brutal" picture of that famous late-Cretaceous event comes from a new census of marine fossils that he conducted with David Raup, SB'53, the Sewell L. Avery distinguished service professor emeritus. The Chicago paleontologists compared survival rates of organisms from that period to see if any earlier adaptations -- in body size or feeding habits, for example -- helped them to survive the catastrophe. Surprisingly, nothing did.
"There really was nowhere to hide," says Jablonski, a professor of geophysical sciences. "What we ordinarily think of as mechanisms for survival just didn't work."
Because of the study's size and subjects -- 350 lineages of marine mollusks, whose fossils are abundant and widespread -- its results provide a good yardstick for how other creatures fared. Organisms with a better chance for survival were those which, like mollusks, were distributed over many continents. That finding is consistent, he and Raup believe, with evidence that dinosaurs as a group were widespread, but specific dinosaur species were not.
Their study "doesn't tell us whether it was an asteroid or a volcano" that caused the extinction, says Jablonski, "but it does constrain some of the killing mechanisms." For example, theories that blame a die-off among plankton don't square with the fact that plankton-feeding mollusks fared no worse than their bottom-feeding cousins.
Another clue to the nature of the "killing mechanisms" comes from geophysical-sciences graduate student Paul Markwick, who has surveyed late-Cretaceous crocodile fossils as part of the U of C's Paleogeographic Atlas Project. Because these cold-blooded animals require mild temperatures, Markwick's finding -- that crocs remained numerous and widespread after the mass extinction -- disputes the notion that a severe climate change killed the dinosaurs. "The cooling may have had a part," says Markwick, "but it clearly wasn't the major player."
So how could cooling fit in? The mechanisms behind two popular theories, an asteroid impact and volcanism, both include a short-term cooling, and the source of that cooling -- blocked sunlight -- would also stop photosynthesis, causing a breakdown in the food chain, a breakdown that would hit dinosaurs hard. A cold-blooded crocodile, says Markwick, "could go kill a wildebeest, and that would keep it happy for the whole year," while a presumably warm-blooded dinosaur would need a constant source of food.
Then the dinosaurs starved to death? "It's not that neat and clear," he says. Other warm-blooded animals survived the extinction -- namely mammals.
Enter a new star suspect. David Schramm believes that the asteroid-impact theory, proposed in 1980 by a group including Luis Alvarez, SB'32, SM'34, PhD'36, is by now "rather nicely established." But he doubts if it can explain other mass extinctions, especially given estimates of the late- Cretaceous asteroid's unusually huge size: 5 kilometers across. A cosmologist, Schramm wondered if other astronomical events could be among the culprits. Last year, he and John Ellis of the European Organization for Nuclear Research updated a theory from the 1970s: Was a nearby supernova to blame?
Schramm, the Louis Block professor in physical sciences, says radiation from an exploding star could destroy the Earth's ozone layer for hundreds of years, leaving phytoplankton and plants unprotected against ultraviolet light. Then "our own sun does the killing" -- as if the ozone hole now over Antarctica suddenly expanded to cover the globe -- and the food chain crumbles from the bottom up.
Using new data on nearby supernovae, Schramm and Ellis also note an interesting coincidence between their best guess for local supernovae rates -- one event every few hundred million years -- and the age of the granddaddy of all mass extinctions: the one that ended the Permian period some 185 million years before the dinosaurs bit the late-Cretaceous dust.
Depending on your view, it's scary or amazing or both. And genetics will soon become a lot more so. Imagine, for example, an amniocentesis test that screens for dozens of genetic disorders -- instead of a mere handful. Laying a foundation for this coming knowledge boom is the mammoth Human Genome Project, biomedicine's $3-billion, ongoing effort to chart the location and chemical sequence of every single gene on the human chromosomes. Amid talk of the project's wonderful progress, some geneticists worry that a discussion of its ethical implications is getting short shrift.
Mary Mahowald's concerns started when she noticed how official descriptions of the Human Genome Project made careful use of gender-neutral language. A medical ethicist and professor of obstetrics and gynecology at the U of C, she wondered, would the project's impact really be so impartial?
Couples, Mahowald points out, don't undergo amniocentesis, abortion, or prenatal treatment of a fetus. Women do. And when a child is born with a genetic condition, a woman is usually the caregiver. In the context of these differences, she says, "gender-neutral language is a mistake. It masks the fact that real differences result in real inequities or disadvantages."
Differences related to reproduction and caregiving, she believes, are crucial areas for understanding the impact of the Human Genome Project. A notion of "gender justice," Mahowald said at a spring conference she organized on "The New Genetics: Implications for Women," may help society negotiate a thicket of new genetic dilemmas without turning physical differences into inequities. Because those differences make complete equality impossible, she says, gender justice instead means making fairness the goal.
Before a Swift Hall audience, she explained the idea through examples like recessive genetic disorders -- sickle-cell anemia and cystic fibrosis are among the most common -- which can affect children only when both parents are carriers. What if a pregnant mother, fearing the presence of such a disorder, asks for a fetal diagnosis when her partner has refused to take the blood test to see if he's a carrier? That means subjecting the fetus to the risks of an amniocentesis -- unnecessarily, if the father doesn't carry the gene. If the test goes ahead, a positive diagnosis would confirm that both parents are carriers -- notifying the father who had preferred not to know.
As part of a three-year study she led on women and the Human Genome Project, Mahowald, assistant director of the University's MacLean Center for Clinical Medical Ethics, asked other genetics experts whether they would allow the amniocentesis in such a case. Some wouldn't, believing that it could undermine the family structure, while others, arguing that a woman's autonomy extends to the fetus, thought amniocentesis should be allowed.
From the viewpoint of gender justice, Mahowald argued, the latter was correct: Health-care professionals should permit the test, because -- while the decision would treat women and men unequally -- the test's burden and risks fall more on the woman than on her partner.
Some gender inequities, of course, are socially created. In the future, Mahowald warned, insurers might prefer to provide differential coverage for women depending on their predisposition to pass on particular abnormalities to their children, as revealed by genetic diagnosis. Important new services -- like a test for the recently identified gene for hereditary breast cancer -- might be covered by private health insurance if a woman's family history indicated an elevated risk, but not covered by public aid.
"Gender justice," Mahowald concluded, "means identifying gender differences and taking account of them, so as to maximize equality." If her predominantly female audience was any indication, society may not yet be equally attuned to that need.
Third Time's the Charm. Millions of children take Ritalin to treat
attention deficit hyperactivity disorder (ADHD), yet surprisingly little
research has considered the drug's optimal dosing. Now, a study by Mark
Stein of the U of C Medical Center shows that three daily doses
of Ritalin -- not the usual two -- may substantially reduce ADHD symptoms, such
as difficulty concentrating or impulsive behavior. Stein found little
evidence to support pediatricians' fears concerning a third, afternoon
dose: that side effects like insomnia or irritability would increase in
incidence. In fact, the only side effect that worsened was a slight loss in
appetite. Though his study was small, its results, says Stein, suggest
prescribing a third dose if a child's symptoms are severe enough to disrupt
late-afternoon activities such as homework, chores, and play with friends
or family.
Written and compiled by Andrew Campbell.